Dave Williams is the senior principal of PAE engineers, with 20 years of experience in mechanical engineering. Through extensive collaboration with data centers, laboratories and healthcare organizations, he has become very skilled in providing design for controlled environments. Dave focuses on reducing operational and maintenance costs and improving energy efficiency. He has considered Kaiser Permanente, Legacy, Tuality and PeaceHealth among his long-term clients. Dave is a LEED-accredited professional who has worked on many sustainable projects, including several LEED and Net-Zero Energy certified buildings.
PAE engineers specialize in designing energy-efficient and high-performance data centers that prioritize sustainability and operational efficiency. Their work integrates innovative cooling solutions and advanced engineering to optimize performance while reducing environmental impact.
Can you take us back to your first experience in designing a data center? What is a project, what are your biggest challenges or learning moments?
The first major is memorable and honestly one of my favorites. This is to convert an old semiconductor facility into a data center. As you can imagine, this conversion faces multiple challenges, from repurposing the device to different operating conditions, increasing resilience and ensuring proper operation when the device is reused is challenging. Debugging is also challenging and ensures that all components, especially cooling equipment and controllers, can quickly increase power failures, thus bringing some interesting and interesting challenges to the solution. Finally, everything is good and the team really stays in touch with many players in collaboration to execute and hand over a good working data center facility.
How have data center design priorities changed over the past 20 years? What are the biggest changes you see in cooling, power generation, and redundancy?
Priorities have changed a lot. The existence of a data center provides a flexible, high-start environment for the server. As the servers become more powerful and adaptable to the environment, facilities have changed. They have become very effective, converting energy-efficient envelopes from converting power to racks and transporting, while HVAC systems provide a stable environment for facilities to transform from large power consumption equipment to innovative, efficient and efficient systems that allow for much lower PUE than ever before.
What is the biggest challenge in designing a data center now? How do factors such as power availability, cooling capacity and stricter regulations affect design decisions?
The availability of reliable power is often the biggest factor in positioning a data center. Availability on land, proximity to fiber, safety and climate focus all focus on all factors, but finding power seems to be a big driving force.
How do PAE engineers integrate sustainability into data center projects from the initial planning phase?
From the end, it’s always a good reminder. Sustainability is a broad term, not just energy, but with the help of the energy required, it usually attracts the greatest attention. Setting low-key goals and sticking to them in the design process is guiding light and helping make decisions. For example, the goal of a data center is to power the server. Any other energy used in the facility supports this and is where we focus on reducing it. The cooling system is the main criminal among them, where we really want to start from the beginning. Given the climate, the facility will be located and optimized in the system to maximize free cooling and considering the approach of avoiding the use of large compressor motorized cooling systems does contribute to efficiency. Subsequently, in the Data Hall, water reduction measures, control optimization, and even consideration of considerations such as renewable energy and rainwater capture, the review of thermal recovery is other items considered when high sustainability is achieved in the facility.
What is the most promising innovation in energy and water efficiency in today’s data centers?
As servers become more resilient and able to handle higher temperatures, the ability to freely cool up to increase time in the year really has an impact. Therefore, many of the data centers deployed now have some of the most innovative and energy-efficient systems installed. Liquid cooling and denser capacity servers present new challenges and opportunities as water has greater potential for large-scale heat transfer than air.
How do you balance high-performance computing needs with sustainability goals? Are there any trade-offs when reducing PUE (power usage efficiency) or water use while maintaining reliability?
If done correctly. The infrastructure must support the server rack requirements through resilient deployment, but if thoughtful, it can be optimized to have very effective functionality, thereby reducing PUE and water consumption without compromising the performance requirements of the compute.
What challenges do you face when implementing renewable energy, advanced cooling or other sustainability measures in high-density data centers?
Challenges vary by project. Sometimes, local regulations or jurisdictional barriers may need to be overcome, and land may not be able to deploy some type of system based on the cost of implementing innovative designs, or to deploy land honestly. It works well for one climate or site and may not be the best choice for another, so this should be considered for every project.
How do AI workloads change data center design? What adjustments are required for power and cooling to support high-density AI racks?
Densification can greatly affect power and cooling systems. All of the larger feeds, circuit breakers, pipes, etc. that support more local power loads will pose challenges, while also having the opportunity to rethink how infrastructure should be laid out, positioned and deployed in a safe, resilient and functional way.
What role do you think AI plays in optimizing energy efficiency and real-time operations in data centers?
It is huge. The ability to create machine learning facilities that use large-scale computing power to optimize systems that optimize energy consumption, which are not all data centers but are the environments built in general. Peak shaving, understanding how each facility operates and how it is optimized in a specific situation is a huge opportunity. Rolling the calculations of multiple facilities and adjust the adjustments to different climate regions. I can only imagine how much AI understands the operation. The suggestions and optimizations it can deploy will have a huge impact on how to adjust and adapt to systems for these facilities.
What is the next generation of AI-powered data centers like? Do you expect extensive use of immersive cooling, on-site renewable energy or fully automated facilities?
Continuing higher density shelves is the trend that may last for several years. This densification creates interesting opportunities for the physical scale of data centers with potentially smaller footprints, but also needs to consider new things needed about infrastructure such as larger circuit breakers/pipes, etc. Immersive cooling is fun and has been discussed in the industry for a while. At the moment, it is not that common and not much use, but it may become more feasible as the density increases. On-site renewable energy is always a good idea for any project, whether a data center is a good idea, but again, when deploying and choosing which renewable energy type and how many types should be installed, the cost, ROI, feasibility, climate and more needs to be taken into account.
Thank you for your excellent interview and hopefully learn more readers should visit PAE Engineer.
